Langchain representation model creates the same representations for all topics (instead of different representations for each topic)

Nobody has claimed this yet.

### Have you searched existing issues? 🔎

- [x] I have searched and found no existing issues

### Desribe the bug

I'm currently working with the Langchain representation model for a project where I need to generate distinct representations for various topics. However, I've encountered an issue where the model produces the same representation for every topic.

```
class LangchainRepresentation(BaseRepresentation):
def __init__(
self,
chain,
prompt: PromptTemplate,
nr_docs: int = 4,
diversity: float = None,
doc_length: int = None,
tokenizer: Union[str, Callable] = None,
chain_config=None,
):
self.chain = chain
self.prompt = PromptTemplate.from_template(prompt)
self.default_prompt_ = DEFAULT_PROMPT
self.chain_config = chain_config
self.nr_docs = nr_docs
self.diversity = diversity
self.doc_length = doc_length
self.tokenizer = tokenizer

def extract_topics(
self,
topic_model,
documents: pd.DataFrame,
c_tf_idf: csr_matrix,
topics: Mapping[str, List[Tuple[str, float]]],
) -> Mapping[str, List[Tuple[str, int]]]:
"""Extract topics.

Arguments:
topic_model: A BERTopic model
documents: All input documents
c_tf_idf: The topic c-TF-IDF representation
topics: The candidate topics as calculated with c-TF-IDF

Returns:
updated_topics: Updated topic representations
"""

# Extract the top 4 representative documents per topic
repr_docs_mappings, _, _, _ = topic_model._extract_representative_docs(
c_tf_idf=c_tf_idf,
documents=documents,
topics=topics,
nr_samples=500,
nr_repr_docs=self.nr_docs,
diversity=self.diversity,
)
# Generate label using langchain's batch functionality, removed truncate documents
chain_docs: List[List[Document]] = [
[
Document(doc)
for doc in docs
]
for docs in repr_docs_mappings.values()
]

# `self.chain` must take `input_documents` and `question` as input keys
# Use a custom prompt that leverages keywords, using the tag: [KEYWORDS]

if "keywords" in set(self.prompt.input_variables):
keywords_list = []
for topic in topics:
keywords = list(zip(*topics[topic]))[0]
keywords_list.append(','.join(keywords))

inputs = [{"input_document": docs, "keywords": keywords} for docs, prompt in zip(chain_docs, keywords_list)]

else:
inputs = [{"input_document": docs,} for docs in chain_docs]

# `self.chain` must return a dict with an `output_text` key
# same output key as the `StuffDocumentsChain` returned by `load_qa_chain`
chain = self.prompt | self.chain
outputs = chain.batch(inputs=inputs,)
labels = [output.content.strip() for output in outputs]
updated_topics = {
topic: [(label, 1)] for topic,label in zip(repr_docs_mappings.keys(), labels)
}

print(f"updated_topics: {updated_topics}")

return updated_topics

langchain_representation_model = LangchainRepresentation(
chain=llm, nr_docs=10,
prompt="""
#Keywords#
{keywords}

#Task#
Based on the provided information filed to a company, derive a label that encapsulates the essence of the information .

#Output Requirement#
- The label should be an impactful phrase that represents the ##derived Summary##.
- The label should also be short and precise.
- Only output the label.
""")

mmr = MaximalMarginalRelevance(diversity=0.7)

topic_model = BERTopic(
min_topic_size=10,
zeroshot_min_similarity=0.7,
ctfidf_model=ClassTfidfTransformer(bm25_weighting=True,reduce_frequent_words=True),
umap_model=umap_model,
vectorizer_model=CountVectorizer(ngram_range=(1,3), stop_words="english"),
representation_model=[mmr, langchain_representation_model],
verbose=True)
```

### Reproduction

```python
from bertopic import BERTopic

```

### BERTopic Version

0.16.4

Start with LangchainRepresentation.extract_topics and the BERTopic representation_model configuration shown in the report. Inspect the representative-document mappings, generated batch inputs, and returned outputs for multiple topics. Done means the representation model produces distinct topic representations where the input topics require them.

Written by the indexing model from the issue text.